▲ NexaFlow Logistics

AI-powered supply chain platform: Agents + MCP + RAG in ASP.NET Core 9

📄 Knowledge Base

6

Supply chain documents indexed and searchable via TF-IDF + dense embeddings

Ask a Question
🔧 MCP Tools

9

Tools available: inventory lookup, supplier search, shipment tracking, code interpreter + more

Browse Tools
🧠 Agent Memory

0

Agent memory entries with recency decay + importance scoring

View Memory
🆕 Features
  • ✓ RAG Chat — TF-IDF + dense vector search over supply chain docs
  • ✓ ReAct Agent — Think-Act-Observe loop with MCP tool calls
  • ✓ Multi-Agent Pipeline — Research Agent → Summary Agent via SignalR
  • ✓ MCP Protocol — HTTP tool discovery + invocation at /api/mcp/
  • ✓ Supply Chain Tools — Inventory, Suppliers, Shipment Tracking
  • ✓ Code Interpreter — Sandboxed C# via Roslyn CSharpScript
  • ✓ RAG Evaluation — Precision@K, Recall@K, MRR, Hit Rate
  • ✓ Agent Memory — Recency + importance scoring
  • ✓ Zero API keys — Full SimulationMode for instant demo
🚀 Quick Start
  1. RAG Chat — Ask questions about NexaFlow policies, suppliers, and logistics
  2. Agent — Watch the ReAct loop: think → call tools → observe → answer
  3. Pipeline — Multi-agent Research + Summary pipeline with real-time streaming
  4. MCP Tools — Invoke inventory, supplier, shipment tools directly
  5. Upload — Add documents to the knowledge base
  6. Evaluation — Run RAG evaluation metrics
  7. Memory — Manage agent memory entries